In the
course of trying to choose the optimal antenna we often find ourselves
comparing manufacturer's specifications about gain, which are all too often
sparse or expressed in inconsistant units of measure. Or worse, relying
only on marketing bullets claiming "best deep fringe".
An alternative approach is to simply compare measured signal levels of one
antenna to another. If we do this on a number of different antennas, let's
call 'em A, B, C, and D, and we find that A is better than B, and B is better
than C, then it follows that A is also better than D.
So the links below provide antenna comparisons of relative field performance of a few popular VHF and UHF antennas.
Some "relative" Measurement Discussion
Consider the goal of finding the tallest person in a class. If we knew each persons exact height, measured precisely in a common unit of measure such as inches or centimeters, we could easily identify the largest height from a form or chart. But what if some were measured in inches, some in centimeters, some in "hands", or some were measured with shoes on, and some with shoes off? If we only had the written information the choice we make could easily be the wrong one.
But if we actually had the students themselves available we wouldn't need to measure. We could just stand them next to each other and visually identify the tallest; even if we had them available only two at at time! As an example consider Alice, Billy, and Jane. If we stand Alice and Billy next to each other and Billy is shorter, and separetely we stand Billy next to Jane and Jane is shorter, then it follows that Jane is also shorter than Alice, and thus Alice is the tallest.
Likewise in the comparisons here, various pairs of antennas are
compared to each other. These are comparisons of relative antenna performance. It actually doesn't matter what the units of measure are, or even whether the measurement device is highly calibrated. We only need to use the same device and with identical condtions to discover whether the signal level received by one antenna is greater or less than the signal received by the other.
Relative and real world field measurements therefore
remove numerous metrics typically found in various manufacturer's specifications. Since such published specifications do not adhere to any industry or organizational standards one can easily be comparing apples to oranges. I.e., what were differences in the antenna
range(s) at which the measurements were made? How accurately were all losses
of baluns, cables, etc., taken into account? How carefully was the
manufacturer's instrument calibrated? What was the instrument's precision? And, in particular, what units were cited gain numbers
expressed in, (dBd or dBi)?
For the record:
- All measurements of antenna pairs here are made back to back, as quickly as possible, in order to
minimize the error contribution from time variant environmental conditions.
- The receive transmission path consists of about 12 feet of Belden 7916A RG6 as
down lead, a Kitz Technologies (1dB NF) amplifier whose gain was measured to be
20dB and flat as a pancake across the combined VHF/UHF spectrum, followed by
another 8 feet of RG6 to the spectrum analyzer.
- In all cases, antennas were tested with their stock baluns, whether built in
or supplied as a 300 to 75 ohm matching transformer.
- Height was about 15' above ground level.
- From measurement to measurement, the antenna under test was rotated to produce
maximum amplitude across all channels within the bandwidth span (not necessarily maximum channel flatness
of any one channel).
- Also from one antenna comparison to another the vertical scale of the analyzer may
change; e.g., 10 dB / division or 5 dB / division. But it is
not changed between measurements of antennas within a comparison
set.
In other words, a lot of care was taken to insure that the predominant variable between comparisons was
the antenna itself.
And A Few More Technical Notes About The Measurements
The following notes attempt to clarify the
measurement environment, conditions, and definitions.
-
Nothing here should be interpreted as attempting to represent an antenna's exact "performance", or gain, as expressed in dB (Decibels) or dBi (dB isotrophic and as commonly found in manufacturer's specifications. The primary metric being compared here is the measured level of a received signal from one antenna, a function of it's sensitivity, or gain, relative to the level received by another antenna. In practice there are numerous other
factors that affect performance; height above ground, reflections that
can cause multipath interference, the quality (and thus loss) of downlead cable, poor house
distribution, splitters, amplifier quality, installation anomolies, and so on.
- None of the captures are intended to present actual (absolute) field strength.
Again, each result is presented as a comparison between two antennas or
configurations, at an identical field location.
- Every effort was made to insure the only variable was the unit under test.
Antennas were raised at exactly the same spot, were located at exactly the same
height (with one qualifier as follows), were carefully positioned, the same
cables, amplifier, etc. were used, and the elapsed time between measurements was
minimized as much as humanly possible. The qualifer is that when comparing
units of significantly different vertical dimensions, such as a Yagi to an
N-Bay, the height chosen on the mast was the lower of the maximum acheivable
heights of the vertical centers of the two designs. To clarify, if
the maximum attainable height of the vertical center of an 8-Bay was 14" below
the top of the mast (because of 8-Bay top mounting bracket position), the vertical center of a compared Yagi would also be
located at 14" below the top of mast (ignoring the possibility that the Yagi
could have been raised another 14"). Nevertheless, all of these physical
aspects represent possible sources of error.
- The Spectrum Analyzer being used is an 80's vintage HP8590B. While it
has a very convenienet 75 Ohm input
impedance, it's
calibration is out of date, and it has some irritating frequency
drift over time. This occasionally causes traces to not line up precisely
across the horizontal axis. However, observations of indicated level while monitoring
stable level signal generators indicate very little gain drift; i.e., the gain is
stable over time. Again, for
these relative measurements, accurate determination of the absolute level
or frequency is unimportant. The important factor is difference in
observed levels (e.g. the level from one antenna relative to the other),
and the internal gain of the instrument does not significantly change over the
typically 15 minutes between antenna measurements. [One more analogy on this topic: imagine the cheapest voltmeter you can buy whose absolute accuracy is only + or - 10%. If you measure one car battery and the needle lands on "10", the actual voltage is anywhere between 9 and 11 volts. But if you measure another battery and the needle lands on "12", it is visually obvious that the 2nd battery has a higher voltage.]
- The measurements are made at the coordinates 39° 21’ 14.63” N 120° 5' 19.54" W.
The terrain is extremely mountaineous and rugged. Note that while
TVFool predictions serve well in LOS scenarios, at these
coordinates they're really more of a hint at reality; are not generally as good
as RFProfiler, and neither tool is highly precise for
this 1Edge and 2Edge, rugged mountain terrain location.
- Channels available at the measurement location include VHF channels 7, 9,
13, and UHF 23 and 26 at a compass heading of 103 degrees (East by South East), and
UHF 15, 20, 24, 30, 33, 44, 47, 51, and 55 at a heading of about 28 degrees
(North East). [Note: following the June 2009 cutover 9 moved down 8.
Thus some VHF comparisons may show different channels].
- Almost all signals are fairly weak in the area, with the exceptions of VHF 7, 8/9, 13 and UHF 44,
and suffer from multipath
distortion due to mountainous terrain. But to arrive purely at a conclusion about
relative received signal strength between two antennas, signal
quality (flatness) is a "don't care". E.g., it doesn't matter that a given channel has nulls (dips)
or peaks across its
approximately 6MHz channel bandwidth. The only thing being measured here
is the relative receive level of one antenna compared to another.
- Where the channels 23, 24, and 26 are involved in the comparisons, side lobe
response can yield differences in receive levels that are dramatic. Take
care as you study comparisons of these channels to take into account that 23 and
26 are about 75 degrees to 24. It means that while looking at captures including channels 23 and/or 26, for example, anything observed for 24 is
irrelevant.
- In all pictures, the images of the separate screen captures of two
antennas are merged. The Green color is the antenna judged to be the
overall better performer. The pictures have 96dpi resolution and if you
have a scroll wheel on your mouse you can scroll up/down to zoom in/out on them.
- At the time some of the earliest measurements were performed UHF Channel 23
was not on the air, and UHF Channel 55 was. After the June 2009 DTV
cutover this reversed; e.g., 23 came on the air and 55 went off.
- At different points in time different vertical scale values were used.
In earlier
measurements 10 dB per division was used, and thus each minor division
represents 2 dB. But in later measurements 5 dB per major division was used, and thus each
minor division is 1 dB. So you may have to squint a little bit on some. The
indication of scale appears at the top left of each capture, as something like
"LOG 10dB/" indicating 10 dB / division.
- The "REF" (vertical axix reference level) also varies somewhat arbitrarily. It is
irrelevant. What ever value was convenient for the channels being observed
was used, and care was taken to insure that absolutely the same level was used
for both antennas being compared.